The primary aim of this project is to understand how gene regulation generates precise spatial patterns in embryonic development. Biochemical reactions frequently operate at concentrations of hundreds to thousands of molecules in a nucleus or cell. Statistical expectation is that such low numbers should display relatively high variability. Yet gene expression patterns display high-precision boundaries, necessary for proper embryogenesis. This study will extend current understanding of noise control in temporal gene expression to the case of spatial patterning in complex eukaryotic development: in particular, anteroposterior body segmentation in the fruit fly Drosophila melanogaster. Using this genetically very well characterized system will contribute to understanding and controlling noise robustness in human development. Noise in segmentation is not well characterized: experimental evidence to quantify expression variability has not been gathered for statistical numbers of embryos; and mathematical models have been descriptive or deterministic, lacking noise terms. This project focuses directly on filling in these experimental and theoretical gaps. In preliminary work, expression variability has been measured both within single embryos and between embryos (work is needed to quantify the specific contributions from molecular noise and from other factors, e.g. embryo size or shape). The proposed work introduces a method for extracting quantitative gene kinetics from measured variability, i.e. using noise as a probe into the underlying biochemistry. Putative kinetic mechanisms will be tested, via stochastic simulation, for the ability to reproduce experimental variability. Parameters will be estimated by reverse-fit optimizations to data (mean and variance of expression levels). Such a mechanism can be used to determine relative contributions to patterning by different types of molecular transport (active vs. diffusive), and for studying the kinetic basis of noise suppression in developmental genetics.